Ozgur Gurcan
Backend Engineer
- Izmir, Türkiye · Open to Remote
- [email protected]
- linkedin.com/in/zgrgrcn
- github.com/zgrgrcn
Summary
Backend engineer with 6+ years of experience building high-traffic, event-driven systems in Java and Spring Boot. Most recently at Trendyol – a platform serving 30M+ customers and 250K+ sellers across 27 countries – I worked across a coupon domain of 38 applications on the checkout critical path, built on Kafka, Couchbase, Elasticsearch and PostgreSQL. I have led architecture redesigns and database migrations end to end (ADR, phased rollout, rollback, post-deploy monitoring) and invest in code quality, observability and developer-productivity automation.
Technical Skills
- Languages
- Java 25, JavaScript, TypeScript, SQL; working knowledge of Go
- Frameworks
- Spring Boot 4, Quarkus, Vert.x, NestJS, Hapi
- Testing
- JUnit 5, Mockito, Testcontainers, WireMock, Cucumber (BDD end-to-end), Pitest mutation testing, JaCoCo
- Data
- Couchbase (XDCR, CBES), Elasticsearch, PostgreSQL, MongoDB, BigQuery
- Messaging & Design
- Apache Kafka, transactional outbox, idempotent consumers, saga/compensation, DDD, hexagonal architecture, CQRS
- Resilience & Performance
- Resilience4j (circuit breaker, retry, bulkhead), Caffeine caching, rate limiting, index and query tuning, virtual threads, API versioning
- Infrastructure
- Kubernetes, OAM manifests, Docker, GitLab CI, AWS, Maven, Gradle
- Observability & Quality
- Prometheus, Grafana, Kibana, OpenTelemetry, Elastic APM, SonarQube, alerting
- Practices
- System design, ADR/RFC authoring, multi-datacenter deployment, on-call and incident response, mentoring
- Spoken
- Turkish (native), English (professional working proficiency)
Experience
Trendyol
Backend Engineer, Coupon domain
to Izmir, Türkiye
- Worked across a domain of 38 applications – Java/Spring Boot and Quarkus services, Go consumers and Kafka connectors – on the checkout critical path, running active-active across two datacenters on an event-driven Couchbase, Elasticsearch and PostgreSQL architecture, serving roughly 4M requests per minute at steady state, close to 7M at peak, and around 800K coupon redemptions a day.
- Authored the architecture decision record for the domain’s core data model and designed the redesign behind it: embedding related entities into a single document to remove an N+1 lookup costing up to 200 extra database reads per request across 11 endpoints. Shipped behind a feature flag with an atomic unit-of-work and rollback. At full rollout it retires around 300M documents, 245 GB RAM and 780 GB disk.
- Built one of the domain’s services from the ground up and later led its migration from Couchbase to PostgreSQL with zero downtime and zero data loss – dual-write, automated consistency checks and a controlled traffic switch with rollback – introducing a repository abstraction that became the team’s reusable migration pattern.
- Was the top contributor to the real-time targeting service: instant coupons triggered by basket, browsing and favourite events plus dynamic marketing cohorts, built on a transactional outbox and a state machine. Elsewhere designed the compensation flow that keeps cross-border coupon creation consistent as a saga.
- Found through BigQuery analysis over a 32 TB coupon dataset that coupons were distributed uniformly regardless of discount value, and designed a quality-based dynamic limit algorithm: high-value coupons reached more customers, wasted generation dropped and system load improved. The same analysis exposed countries receiving zero coupons, traced to an upstream script bug.
- Prepared the domain for 5–10x campaign traffic peaks – capacity planning, load and chaos testing, caching, rate limiting and index tuning – holding p99 at 63 ms on the main read path. Ran multi-datacenter deployments through the platform’s deployment manifests: pod sizing, autoscaling, progressive rollouts, active-zone shutdown to shift traffic between DCs, and a cost scale-down that zeroed idle replicas across every tier.
- Drove the domain’s end-to-end BDD suite to 340 scenarios covering nine services, running on every merge request with parallel execution and a fail-only rerun stage that separates flakes from real regressions. Moved 16 services toward SonarQube Grade A, led the Java 25 and Spring Boot 4 upgrade, and carried the on-call rotation, leading incident response and authoring RCAs.
Getir
Software Developer
to Izmir, Türkiye
- Owned the gateway service behind Getir’s mobile and web applications – dashboards, shop detail views and product navigation – built on Node.js (NestJS, Hapi) and MongoDB, directly shaping the customer-facing experience.
- Designed and shipped a new service centralizing shared logic and downstream service calls, eliminating code duplication, improving performance by over 15% and simplifying error tracing across the platform.
- Introduced team-level engineering processes – test quality metrics, documentation standards, structured onboarding and an improved production-defect workflow – contributing across both Node.js and Java/Spring Boot services.
PiA (People in Action)
Software Developer
to Remote, Türkiye
- Migrated Vodafone Spain’s platform to a RESTful microservice architecture, improving system efficiency.
- Built the backend infrastructure for Vodafone Albania from the ground up, delivering a RESTful API for customer support operations.
- Mentored junior engineers, providing technical guidance and owning task breakdown.
Compelis
Software Engineer – early-stage startup, part-time
to Remote, USA
- Built competitor-intelligence software for the e-commerce sector on a serverless architecture.
Corebit Tech
Backend Developer
to Izmir, Türkiye
- Developed, debugged and shipped web application code while completing my final year of university.
Aselsan
Software Developer (Intern)
to Ankara, Türkiye
- Built pre-compile syntax and logic checks using Abstract Syntax Trees, regular expressions and C++ to enforce code quality across company software components.
Engineering Initiatives
AI Code Review Bot
(MCP servers, vector storage, Slack, n8n) – reads the codebase, documentation and Jira tasks to generate contextual MR comments, and answers system questions directly in Slack.Documentation Automation
(n8n, GitLab) – diffs merged code changes against wiki standards and regenerates technical documentation automatically, removing recurring manual upkeep from the team.Self-Service Config Pipeline
(GitLab CI) – replaced a manual, code-change-driven process with a validated pipeline the client team runs itself, cutting update time from 30 minutes to under 30 seconds.Load-Test Data Generator
– a single parallel call replaces around 30 manual requests and generates tens of millions of coupons, cutting load- and chaos-test setup from an hour to under five minutes.
Education
Izmir Katip Celebi University
MS, Computer Software Engineering (GPA 4.00/4.00)
to
Dokuz Eylul University
BE, Computer Engineering (GPA 3.33/4.00)
to